Judge Amit Mehta will shortly provide his remedy to Google’s monopoly in internet search. Fiona Scott Morton and Paul Heidhues argue that the remedy must include a cap on Google’s payments to the mobile phone manufacturers, carriers, and web browsers that propelled its monopoly. Because any outright ban risks harming Google’s current partners in the short term, Judge Mehta should consider pursuing a flexible ban that instead limits the revenue these partners can receive from Google in order to encourage market entry and competition.
Judge Amit Mehta is expected to announce within weeks the remedy to Google’s internet search monopoly. During the remedies trial phase of the Department of Justice’s case against Google, which concluded in late May, Judge Mehta homed in on the key problem in designing a remedy. Central to Google’s anticompetitive conduct is its erstwhile strategy of paying Android mobile phone original equipment manufacturers (OEMs), mobile carriers, and proprietors of web browsers like Apple and Mozilla a share of its advertising revenue to exclusively preinstall Google Search and set it as the default search engine. This conduct, however, has created two opposing forces that complicate unwinding Google’s revenue-sharing agreements. The first force is that Google used these payments to keep its rivals out of the search market. The other is that Google’s payments provide important revenue streams to those partners (for a more detailed explanation of this trade-off, see the authors’ co-authored paper here). While this second force does not justify Google’s payments for exclusive and preferred treatment, neither can their importance to Google’s channel partners—the OEMs, carriers, and browsers—be altogether ignored in the remedy. Rather than outright ban Google from paying its current channel partners, Judge Mehta should consider calibrating the deals to prevent Google from continuing to dominate the internet search market, thereby facilitating market entry and long-term competition.
The trade-offs of banning Google Search from entering revenue-sharing deals
Google’s channel partners collectively received massive sums in the billions for refusing to install rival search engines. The resulting lack of competition in search caused by these exclusivity deals, however, has harmed those channel partners in the longer run. If a channel partner, for example an OEM, had faced a choice between four search engines relatively equal in market power, it could have chosen to pre-install the search engine that offered it the most favorable terms. Rather than needing to accept whatever the monopolist offered, as it does now with Google, the OEM could have held a bidding contest between the rival search engines to extract a higher share of internet search advertising revenue. Those higher payments to the OEM would have lowered its costs, which in turn would have helped it compete for end users by offering lower-priced handsets. Search competition would therefore have likely moved resources from the monopolist to the OEM and on to the end users, benefitting both (end users would have paid less for handsets and services, while OEMs would have received a bigger cash subsidy).
Today, Google’s monopoly position allows it to induce channel partners to block entry in search with relatively low payments. Judge Mehta has realized that if he takes away those low payments, it could unblock Google’s channel partners from signing agreements with other search engines and stimulate entry and competition. But at the same time, removing those payments will create temporary harm for the channel partners who today depend on a share of Google’s monopoly profits for their revenue.
When Judge Mehta asked the channel partners (Mozilla, Motorola, etc.) during the remedies trial how banning Google from making these payments would affect their welfare, they testified that a ban would harm them. Judge Mehta wondered if it was acceptable to “harm these markets” in order to fix the search market. The DOJ’s answer was yes, it is possible that a monopolist in one market may be propping up other markets (markets for OEMs, browsers, carriers), but this is not an acceptable basis for protecting those ancillary markets. In general, firms in a market cross-subsidized by a monopolist may lose those subsidies when competition takes hold, but this is not a harm to competition, and therefore is not a policy concern.
In the case of Google Search, however, we do not even need to rely on this argument. The channel partners will benefit from new competition between search engines in the long run. The concern is an issue only in the short run because competition will create gains for all parties except Google in the long run. The moment payments from Google are banned, channel partners will face higher net costs (and perhaps charge higher prices) until rival search engines —perhaps including some new entrants who incorporate artificial intelligence into their service—begin to bid for their traffic. At that point, these channel partners can extract the larger payments that competition generates and will then have lower net costs, which will enable them to charge end users lower prices than they did under the Google monopoly in addition to benefitting from more choice and quality competition. Channel partners, and their consumers, thus gain overall from a ban on Google payments that stimulates entry into the internet search market.
This debate illustrates that consumers have partially conflicting interests over time: while users want browser and handset markets to work and deliver benefits today (which Google payments help a little), they also want competition in search which delivers choice, quality, and financial benefits tomorrow (which Google payments harm a lot). Judge Mehta’s role is clear: establish those competitive conditions for tomorrow. He should seek to prevent Google from making payments that maintain its monopoly. But he can also establish conditions that bring the long-run competition into place as soon as possible by protecting the emergence of business arrangements that promote competition for web browsers and mobile operating systems. To this effect, some payments from Google to its current channel partners in the short term may remain acceptable and ease the transition, as we explain below.
Calibrating revenue-sharing deals to enable long-term competition
The bedrock condition of an effective remedy must be to prohibit Google from entering
contracts that require any channel partner to provide it with exclusive, default, or preferential positions. By contrast, rival search engines are starting from such a low market share that they may enter such contracts and pay for favorable positions, including exclusive, preinstalled, and default positions.
Yet, if a third party chooses for its own reasons to preinstall or otherwise favor Google Search without these conditions, Google should be permitted to pay that channel partner a limited share of the search revenue. The remedy, however, must place three crucial limitations on any such arrangement.
The first condition is that the third parties should retain total freedom to deploy as they see fit the Google functionalities that default to Google Search (including the search widget, Chrome, Maps, etc.). Thus, arrangements with Google would be terminable at will by the third parties, who could cease using Google in whole or in part, at any time, and for any reason (other than those proscribed by law). Google should have no claim or expectation relating to its search engine’s continued placement, nor any claim to revenues generated by such continued placement. And if a channel partner has made Google search a default, consumers must retain the right to easily change the default search engine on their device or browser. (All consumers should be able to easily change their default search engine at all search access points.)
Second, any such arrangement should be subject to a coverage cap: Google should be enjoined from sharing any of its search revenue from usage that exceeds 50 percent of the partner’s distribution channel. By limiting payment to (a random) half of users, the coverage cap gives OEMs and browsers an incentive to go out and find another search engine that will give it revenue for the other half of users. If preinstallation is random, then 50 percent of users will deliver 50% of potential search revenue. We do not recommend allowing channel partners to select high-revenue users and install Google search on their handsets, as a small percentage of users could direct most potential ad revenue. A 50 percent threshold on market share (50 percent of randomly selected users) allows a rival search engine to achieve the same scale of operations as Google. Any larger share for Google would preserve its dominance by definition.
Third, Google should be permitted to share with a channel partner no more than 40 percent of the revenue from any of the consumers in the covered group just discussed. The trial revealed that 40 percent is approximately the share of all search revenue that Apple earns from its deal with Google. The price cap is necessary to stop the obvious workaround of Google offering to pay 80 percent of revenue on half of users and obtaining tacit support from the channel partner to preinstall Google for the other half for no payment. On average Google would serve all the users at a 40% revenue share, just as it does today, thereby perpetuating Google’s monopolistic practices. Our full proposal, which considers other incentives of Google and channel partners, is available here.
Judge Mehta could exempt very small channel partners from a cap on the grounds that if Google signed exclusive agreements with all of them that would still constitute a small enough share of the market to allow entry. Indeed, along these lines Motorola has proposed exempting smartphone manufacturers from a payments ban if they collectively control less than 30% of the smartphone market (see page 13 of Motorola’s amicus). This is a risky direction to take, however, because any one channel partner may be small, but the impact on search entry depends on the cumulative market share of all of them. And because Android handsets are manufactured by a number of different OEMs, the exception proposed by Motorola likely covers all smartphone manufacturers except Apple, thereby making successful entry of competing search engines excessively hard.
Meanwhile, Mozilla has proposed exempting independent browser developers from a payments ban, by which they mean browser developers that do not provide a desktop or mobile device operating system or mobile devices (see page 18 of Mozilla’s amicus). However, this would allow OpenAI’s browser, currently under development, and most potential acquirers of Chrome in the case of a remedial divestiture, to receive payments from Google. Such a proposal would be counterproductive and again might require an aggregate market share cap on exempted partners, which is hard to police in a fast-moving market. Rather than exempting some market participants and not others—which could be arbitrary—Judge Mehta could allow all of them to obtain partial payment, as we propose above.
In addition, Judge Mehta must determine when to impose the payment cap. If a cap on payments from Google is announced immediately and takes place quickly, entry in search will occur relatively rapidly, thus increasing payments from all search competitors, as well as increase choice and quality. Any delay in restricting Google’s payments will chill market entry, not only because of the delay itself, but also because of increased uncertainty about future entry conditions. Delay therefore perpetuates Google’s stranglehold over the search market, perhaps aids it to dominate artificial intelligence, and thereby entrenches its monopoly for another decade or two. By the time a delayed cap takes effect, any possible deals with search rivals may well be worth very little because the market will have tipped.
Participants in the search market and channel partners now have their eyes opened to Google’s strategy. The changing landscape due to AI and the resulting new ways of engaging with information on the internet creates options for new businesses to enter, particularly if Judge Mehta creates an opening for competition to thrive. Entrants, however, need certainty concerning the competitive environment or they cannot raise the capital needed to invest and support robust entry. A court-ordered cap on Google payments for five years, with a review in three years, provides that certainty. In three years, if search rivals’ collective market shares have passed 50 percent, the court could end the payment cap. At that point, if generative AI is the way users search, then payments from generative AI companies will be flowing to channel partners, giving them all the resources they need. And indeed, if generative AI companies are already providing search-like functionality today, this suggests the short-run transition will be quick. Generative AI companies may be ready to ink default deals with distribution partners on the first day of the court-ordered remedy, especially since that day may be delayed for some months or years due to appeals.
Given the current opportunity for innovation in search—and the long delay in enforcement against Google’s anticompetitive conduct—Judge Mehta cannot postpone action on Google’s payments for preferential and default status. Our proposal demonstrates immediate action need not pull the rug out from under channel partners. By capping compensation of any voluntary Google preferential treatment to cover at most 50% of a channel’s search revenue and capping the revenue share percentage at 40%, the court would provide an on-ramp for entrants and a stable transition for channel partners. No less is required to foster competition and innovation in the market.
Authors’ Disclosures:
Fiona Scott Morton is an economic expert in the United Kingdom for a group of advertisers seeking damages from Google and a group of users seeking damages from Meta. Within the last three years, Scott Morton has consulted for Microsoft, Bard, Regeneron, Tapestry, SiriusXM, GM, and several electric vehicle makers. She regularly works as an expert witness for government plaintiffs on matters that are confidential.
Paul Heidhues has engaged in competition consulting in the context of trucking and timber industries in collaboration with E.CA Economics within the last three years. He has also advised E.CA Economics on work for Apple in the context of a competition case.
You can read our disclosure policy here.
Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.
Subscribe here for ProMarket‘s weekly newsletter, Special Interest, to stay up to date on ProMarket‘s coverage of the political economy and other content from the Stigler Center.
